This research focused on cohort studies, although their methodology can be criticized because of sampling techniques, reliance on arrest records rather than on all police contacts, and State variations in criminal codes. Initially, four studies conducted between 1922 and 1955 are reviewed which have provided support for the belief that mental patients are less inclined to criminal activity than other people. The data from these projects demonstrate that mental patients discharged before World War II were less often arrested than were members of the general public. In contrast, eight American studies done since 1965 have found that arrest and conviction rates of former mental patients equaled or exceeded those of the general population in some crime categories. Following summaries of these studies, an analysis of their findings begins by noting that although arrest rates for the general population and mental patients have increased since the 1950's, the rate of acceleration has been greater for the latter group. This trend is attributed to hospital policies which discharge patients after short stays instead of committing them for years and the tendency of the criminal justice system to place offenders with any history of psychiatric problems in mental hospitals instead of jails. These studies have provided conclusively that the small subset of patients who have prior criminal records accounts for a large majority of postdischarge arrests. Persons with diagnoses of personality disorders, alcoholism, and drug dependence have disproportionately higher arrest rates. However, this could be caused in part by demographic factors associated with their distribution in the general population. Traditionally, most arrests of mental patients were thought to have stemmed from minor offenses, but research has indicated that violent crimes or crimes against persons were frequently committed by mental patients who had high arrest rates. The article concludes that patients discharged from mental hospitals are not, by virtue of their psychiatric disorders or hospitalization, more likely to engage in criminal activity than people with similar demographic characteristics who have no history of mental illness. Increases in arrest rates have been caused by patients who had arrest histories prior to hospitalization and were classified as mentally ill by the courts. Footnotes and 28 references are included.